Dortmund, Germany
I get honest data out of machines, then write the software that reads it. CO₂ electrolysis test benches at Fraunhofer. Car infotainment computers at HARMAN. Small devices that shipped and are still running.
What I am building now
Scientists at Fraunhofer were running experiments on a test stand, then working the results out in Excel hours later. I asked if I could fix that. It became my master's thesis.
How I work
Most data problems turn out to be hardware problems in disguise. A noisy bus, a late interrupt, a clock that resets when the power drops.
Fix broken rows, line every sensor up in time, put it somewhere you can ask questions of. A run should be ready to read in minutes.
Spot the moment the behaviour changes. Forecast against what the machine did before. This works because the input was trustworthy.
Selected projects
A.D. Paul & Co.
Reads heart rate and blood oxygen. The sensor fills its own buffer, so the processor never holds it up.
Published · Scopus
A wheelchair that drives out of a hospital bed. Sixteen motor channels and its own battery system.
Vishnu Automation · Sold
Feeds CNC machines. Three different voltages on one board, kept apart so the pumps cannot spoil the readings.
Two were sold to customers. Two are published. All of them had to survive contact with real hardware.